通过删概念再训练，让模型摆脱偏见并提升可解释性。
Debugging Concept Bottleneck Models through Removal and Retraining
- 专家删掉错误概念，再用新标签重训模型。
- 在多个数据集上，准确率显著优于旧方法。
- 适合需要调试模型偏见的领域专家使用。
概念瓶颈模型（CBMs）利用人类可理解的概念预测任务标签，使领域专家不仅能验证模型输出，还能在测试时干预错误概念。然而，此类干预无法解决模型与专家推理之间的系统性偏差问题，例如模型从有偏数据中学习到捷径。为此，我们提出一个通用的可解释性调试框架，包含移除与重训两步：在移除阶段，专家根据概念解释识别并剔除不当概念；在重训阶段，引入名为CBDebug的新方法，利用CBM的可解释性将概念级用户反馈转化为样本级辅助标签，进而实施监督式偏见缓解与针对性增强，降低模型对不当概念的依赖。我们在真实和自动专家反馈下评估该框架，结果显示，相较于先前重训方法，CBDebug在多个CBM架构（PIP-Net、Post-hoc CBM）和含伪相关性的基准上均有显著提升。
原文摘要 · Abstract (English)
Concept Bottleneck Models (CBMs) use a set of human-interpretable concepts to predict the final task label, enabling domain experts to not only validate the CBM's predictions, but also intervene on incorrect concepts at test time. However, these interventions fail to address systemic misalignment between the CBM and the expert's reasoning, such as when the model learns shortcuts from biased data. To address this, we present a general interpretable debugging framework for CBMs that follows a two-step process of Removal and Retraining. In the Removal step, experts use concept explanations to identify and remove any undesired concepts. In the Retraining step, we introduce CBDebug, a novel method that leverages the interpretability of CBMs as a bridge for converting concept-level user feedback into sample-level auxiliary labels. These labels are then used to apply supervised bias mitigation and targeted augmentation, reducing the model's reliance on undesired concepts. We evaluate our framework with both real and automated expert feedback, and find that CBDebug significantly outperforms prior retraining methods across multiple CBM architectures (PIP-Net, Post-hoc CBM) and benchmarks with known spurious correlations.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。